The article under analysis was published on the official site of the information resource "ScienceDaily" written by Imperial College London

https://www.sciencedaily.com/releases/2022/05/220505114646.htm
The article under analysis was published on the official site of the information resource “ScienceDaily” written by Imperial College London. The title of the article is “'Nanomagnetic' computing can provide low-energy AI”. The article was published the 5th of May, 2022.
In this article highlights the recent an important topic in the field of science is the creation of artificial intelligence. In particular, the author attacks the new method with positive criticism, because it can reduce the energy consumption for artificial intelligence, which is currently doubling. Author lays special emphasis on the that energy consumption doubles every 3 months.
The article opens in a general statement that it is possible to create artificial intelligence using tiny nanomagnets that interact like neurons in the brain. Here the author also says the team presented the first proof that nanomagnetic networks can be used to perform processing similar to artificial intelligence. Further on the author talks about the results of this study. The results were positive, nanomagnets can be used for "time series forecasting" tasks, such as predicting and regulating insulin levels in diabetic patients. The author emphasizes that the new method will be an excellent assistant in medicine.
The author goes on to say that artificial intelligence seeks to reproduce the work of parts of the brain where neurons interact with each other to process and store information. According to his opinion it is a burning issue as the team was able to use the magnets themselves to process and store data, getting rid of intermediaries in software modeling and potentially offering huge energy savings.
The author moves on to proceeds to the description of the nanomagnetic state. Moreover, author claims that the application of a magnetic field to a network of nanomagnets changes the state of the magnets depending on the properties of the input field, as well as on the states of the surrounding magnets. After that the author touches upon another burning issue a which is the fact that by increasing the spins into arrays with nano patterns, it turned out to achieve the necessary control and reading to provide information.
To continue his reasoning the author moves on to the fact that training AI to perform even relatively simple tasks can require a huge amount of energy. Namely, most of the energy used to achieve this goal in conventional silicon chip computers is wasted on inefficient electron transfer during processing and storage in memory. Further on, the author claims that the innovation can make nanomagnetic calculations 100,000 times more efficient than conventional calculations. Another fact that is highlighted by the author is the meeting of the is that nanomagnets do not rely on the physical transport of particles such as electrons. Author underlines that in this energy saving is the most important part of science.
In the conclusion the author notes the integration of magnetic systems b into conventional computers to increase energy efficiency when performing intensive computing tasks. The article discusses the positive response to the introduced new method of using artificial intelligence.
I find the article informative, because they can be used in wearable devices for processing biometric data about the body. I can agree with the author that this introduction will help predict and regulate insulin levels in people with diabetes, or the detection of abnormal heartbeat. However, it will be possible not to waste a lot of energy and help people track their health status.
https://www.sciencedaily.com/releases/2022/10/221011160846.htm
The article under analysis was published on the official site of the information resource “ScienceDaily” written by University of Virginia School of Engineering and Applied Science. The title of the article is “Materials science engineers work on new material for computer chips”. The article was published the 31th of October, 2022.
In this article highlights the recent an important topic in the in the field of engineering is reduction of electricity costs for computing equipment. In particular, the author attacks the new method with positive criticism, because since the power grid is already experiencing stress due to weather events. Author lays special emphasis on the that the economy is moving from fossil fuels to renewable energy sources, engineers urgently need to align the energy consumption curve of computing equipment.
The article opens in a general statement that Information, communications and technology account for 5% to 9% of total electricity consumption worldwide. Here the author also says that materials systems that will allow the semiconductor industry to jointly place computing and memory on a single chip are being investigated. The author emphasizes that right now we have a computer chip that performs its computing operations with a small amount of memory.
The author goes on to say that members of John Bielefeld's Multifunctional Thin Film Group contribute. According to his opinion it is a burning issue as when a computer chip wants to communicate with the memory of a larger memory bank, it sends a signal over the line, and this requires energy. The author moves on to proceeds to description of research in the laboratory. Moreover, author claims that research promotes materials with electrical and optical properties that make modern computing and communication possible. After that the author touches upon another burning issue a which is the fact that the material, they are interested in is hafnium oxide, which is used today in the production of cell phones and computers.
To continue his reasoning the author moves on to the fact that the disadvantage is that in its natural state, hafnium oxide is not a ferroelectric. Namely, over the past 11 years, it has become known that hafnium oxide atoms can be manipulated to obtain and retain a ferroelectric phase or structure. After that, the author dwells in detail on the description of Fields' article, which clearly demonstrated methods for measuring very thin films and mechanical stresses. Further on, the author claims that Yashevsky solved this problem by using several different methods to measure oxygen vacancies in the team's thin films and correlated this with ferroelectric properties. Another fact that is highlighted by the author is the meeting of the is that Yashevsky's experiments also show that oxygen vacancies can be one of, if not the most important parameter for stabilizing the ferroelectric phase of a material. Author underlines that the generally accepted opinion is refuted, according to which the size of a crystal called a grain stabilizes hafnium oxide.
In the conclusion the author notes that next step will be to study how the choreography of oxygen atoms in the material promotes awakening and fatigue, which requires studying the dynamic location of vacancies. The article discusses the positive response to the introduced new a method of research that aims to reduce costs.
I find the article informative, because it shows how much scientists are immersed in the study of new substances that can help in computing. I can agree with the author that their research advances the semiconductor industry's ambitions to save energy in computing by co-locating computing and memory elements in an integrated circuit. However, the demand for computing equipment is increasing every year, and it is actively being introduced into modern life.